Re: Creating a SWOP proof with an Epson
Re: Creating a SWOP proof with an Epson
- Subject: Re: Creating a SWOP proof with an Epson
- From: Dave Tuemmler <email@hidden>
- Date: Sun, 16 Jul 2006 11:05:03 -0400
Thanks Terry for the detailed thoughts on a complex issue (for me at
least).
On Jul 16, 2006, at 9:59 AM, Terry Wyse wrote:
I've never quite understood the logic or argument of using a UV-cut
spectro for linearization. I'll include an example at the end of this
post of what I've heard the logic is about and why I think it's
flawed. But I digress...
Dave, as far as you not being able to use your non-filtered Eye-One as
ColorBurst "requires" a UV-cut spectro, the problem is worse than
that. If I'm not mistaken, I believe ColorBurst uses by default a
UV-cut X-Rite Pulse spectro for all their canned linearizations. So
even if you had a UV-cut Eye-One, your linearization would not match
theirs due to the differences in instrumentation. The differences
between instruments is not trivial in my experience. If I had a
working Eye-One here I'd be willing to test this but I don't at the
moment.
It's been a while since we tried using our i1 so I forget the exact
outcome but I do remember it being way off. So if I want to linearize
probably my best bet is to buy a UV Pulse instrument. Same problem,
over $1000 to linearize which might improve results or not.
As to the non-filtered vs. UV-cut theory for linearization....
From what I understand, it has mostly to do with measuring the the
yellow ramp on a paper with a significant amount of optical
brighteners. Let's see if I can explain it adequately...
My understanding as well and at one time explained by Colorburst.
Let's say you have a media that measures L* 97 a* 0 b* -5 with an
un-filtered spectro. We'll focus on the b* value as that's what's
relevant here. When you're measuring the yellow ramp, at 0% you'll get
the full b* -5. As soon as you start adding yellow ink, which has a
+b* value, you'll get to a point where the spectro will measure a
"neutral" value and essentially see the yellow ink as gray. Example:
0% = L97 a0 b-5
1% = L96 a0 b-3
2% = L95 a0 b0 <-----
3% = L94 a0 b+2
4% = L93 a0 b+4
...and so on
You can see, in my bogus example, that at 2% the yellow ink has
effectively neutralized the optical brighteners in the paper causing
it to appear "colorless" or neutral. The argument goes that this will
fool the linearization application and foul up the linearization. In
ColorBurst's case, since they use chroma for linearization, this sort
of makes sense.
So according to this a UV spectro and at least the Pulse is probably
what Colorburst is using, which hopefully I can confirm.
The argument goes that since a UV-cut spectro would filter out or
cancel the effects of optical brighteners, you'd be starting at a
near-zero b* value for paper white and any amount of yellow ink would
result in a +b* value.
But that's only if you calculate your linearization based on absolute
values.
If you take your measurements but then "zero" out on the media (add +3
L* and +5 b*), giving you relative instead of absolute values, it
works just fine....
0% = L100 a0 b0
1% = L99 a0 b+2
2% = L98 a0 b+5 <-----
3% = L97 a0 b+7
4% = L96 a0 b+9
In theory that makes sense but does the software let you change the
values?
So, unless I'm missing something, using relative readings to calculate
chroma (or whatever) would work just fine regardless of spectro
filtration. This would even make the linearization process relatively
immune to the differences in instrumentation. At least in my
observations, various spectros will differ by quite a lot in their
absolute measurements but they are pretty darn close when compared
"relative" to each other. This is also why that, even if you have two
spectros that measure quite differently, if the same spectros are used
to create both the source and destination, the results will look very
similar. But try mixing/matching spectros for source/destination and
you've got a recipe for distaster.
So I need to talk to Colorburst about what spectro they use.
Some might argue that this method cancels out the effects of the paper
itself, rendering the paper as "zero" (L100 a/b0), but in my opinion
(and it's only an opinion) this would be the proper way to LINEARIZE
anyway. In all my years in prepress <slash> printing,
linearization/calibration is ALWAYS done relative to the substrate. If
it was film, you zeroed out on the clear base before measuring dot%,
if it's a plate the background metal is zeroed, on a press you first
zero out your paper. In fact, this is even fundamental to the new "G7"
methodology of press calibration; the press's tone curve (NPDC) is
calculated RELATIVE to the substrate.
My last argument "against" UV-cut spectros is based on visual
comparisons of proofs. In every case I've observed, it certainly
appears to me that the non-filtered spectro ALWAYS produced the
superior visual match as compared to taking the same measurements with
UV-cut filtration. For this reason, I've pretty much settled on
non-filtered measurements when going for critical color proofing
applications.
I will have to look at the software to see if I can adjust the values
using my non-UV i1. However this might take a while due to a heavy
workload before vacation.
Regards,
Terry Wyse
Thanks again for your insight.
Stretch
On Jul 15, 2006, at 11:49 PM, Dave Tuemmler wrote:
We have the lite version of ColorBurst that came with our 4800 Pro.
The software requires a spectro with a UV to linearize. We have an
Eye One but not the UV version & I'm not eager to spend $1500 for
another i1. I'm assuming re-linearization would be a benefit. I
guess the question is how much do you think it will matter. I'm
also thinking of upgrading to the full version so we can load custom
profiles.
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Colorsync-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden